Resume Skill Match

1. System Architecture

• Architecture Pattern: Modern Monolith with Next.js App Router.
• Runtime: Node.js 20+ / Edge Runtime support.
• Concurrency Handling: Utilizes Upstash QStash for background task orchestration and Redis for state caching.
• Type Safety: End-to-end type safety with TypeScript and Zod schema validation.

2. Feature Specifications

A. Multi-Format Document Processing

• Engines:
• PDF: Text layer extraction via pdf-parse.
• Word (DocX): Native XML parsing via mammoth to preserve document structure.
• Images: Optical Character Recognition (OCR) powered by Tesseract.js for PNG/JPG resumes.
• Excel: Data extraction via exceljs.
• Pipeline: Automatic file type detection → Buffer stream processing → Raw text normalization.

B. AI-Driven Analysis Engine

• On-Device Embeddings: Uses @xenova/transformers to run feature-extraction models locally in the browser/server environment. This allows for semantic similarity matching without external API costs (OpenAI-free).
• Skill Taxonomy: A proprietary skill-taxonomy.ts library that handles:
• Alias Matching: Normalizes variations (e.g., "React.js", "ReactJS", "React" all map to "React").
• Canonical Mapping: Categorizes extracted signals into standard industry skill sets.
• Scoring Logic:
• Technical Score: Ratio of matched skills vs. job requirements.
• Semantic Score: Vector distance between the resume embedding and the job description embedding.
• Experience Score: Regex-based extraction of "Years of Experience" signals.

C. Dynamic Recruitment Intelligence

• Readiness Assessment: Categorizes candidates into Ready, Review, or Watchlist based on a weighted composite score.
• Gap Analysis: Automatically identifies the "Top Gap" (most critical missing skill) based on the target role.
• Automated Interview Planning: Generates contextual interview questions based on the candidate's specific skill overlaps and deficiencies.
• Audit Trail: A transparent log of the analysis steps (e.g., "Normalized 15 skill signals," "Computed embedding similarity").

D. UI/UX & Data Visualization

• Interactive Dashboard: Real-time feedback using Zustand for state management (via ui-store.ts).
• Visual Analytics: Radar/Bar charts powered by Recharts to visualize Technical vs. Experience vs. Semantic scores.
• Accessibility: Accessible UI components built on Radix UI primitives and Tailwind CSS 4.0.

3. Security & Infrastructure

• Authentication: NextAuth.js v5 with Role-Based Access Control (RBAC) implemented in rbac.ts.
• Database: Prisma ORM with a focus on relational integrity; easily switchable between SQLite (dev) and PostgreSQL/MySQL (prod).
• Environment Validation: Strict environment variable checking via env.ts to prevent runtime failures.
• Observability: Integrated Sentry for error tracking and performance monitoring.

4. Technical Performance

• Zero-API Embeddings: By using Xenova, the system generates embeddings locally, ensuring data privacy and $0 inference costs.
• Optimized Styling: Tailwind 4.0 "Engine-First" approach for minimal CSS bundle sizes and ultra-fast builds.
• Post-processing: Uses bcryptjs for secure password storage and zod for 100% reliable API payload validation.

Module Stack

Timestamp

Released: May 2026